Paraxanthine

Human trials of paraxanthine itself

The Exercise and Sport Nutrition Lab at Texas A&M University (Richard Kreider) ran three of the trials, and those trials and a fourth from Lindenwood University list coauthors affiliated with the company that supplies the ingredient (PMID: 34836235, PMID: 34960030, PMID: 38725238, PMID: 38745415). None of the five trials below named a single primary endpoint in its abstract, and each tested many outcomes.

• Yoo 2021, double-blind placebo-controlled crossover, 13 healthy adults, single 200 mg dose. Card-sorting errors were lower at 6 hours (p=0.04), as were perseverative errors (p=0.03), a short-term memory reaction time was faster at 2 hours (p=0.03), and vigilance was better sustained at 2 and 4 hours (PMID: 34836235).
• Xing 2021, double-blind placebo-controlled crossover, 12 healthy adults, 50, 100 or 200 mg as a single dose and then daily for 7 days. The main between-treatment test showed only a trend on one memory measure (p=0.06); other differences came from comparing changes from baseline with 95 percent confidence intervals, and were more consistent at 100 and 200 mg (PMID: 34960030).
• Yoo 2024, double-blind crossover, 12 trained runners, placebo, 200 mg caffeine, 200 mg paraxanthine or both, then a 10-km treadmill run. With paraxanthine, card-sorting correct responses rose from before to after the run (6.8 percent, p=0.012), perseverative errors after the run were lower than with caffeine (p=0.026), and reaction time after the run was faster than with placebo (p=0.029); adding caffeine to paraxanthine gave no extra benefit (PMID: 38725238).
• Gross 2024, double-blind placebo-controlled crossover, 21 adults, 100, 200 or 300 mg. The 200 mg dose raised energy expenditure compared with placebo and was associated with lower hunger ratings than placebo; 300 mg raised free fatty acids, and glycerol did not change (PMID: 38745415).
• Bingol Diedhiou 2026, double-blind crossover, 14 male university rowers, 200 mg each of caffeine, paraxanthine, both or placebo before a 2,000-m ergometer trial. Only the combination of caffeine and paraxanthine improved performance over placebo (p=0.044); paraxanthine alone did not show a clear performance benefit, and it was associated with better subjective sleep than the caffeine conditions. The authors list only university affiliations (PMID: 41918248).

Human pharmacology studies

• Lelo 1986, six healthy men given caffeine, paraxanthine, theobromine and theophylline on separate occasions. Paraxanthine's half-life was 3.1 hours and caffeine's 4.1 hours, with similar clearance and volume of distribution (PMID: 3756065).
• Lelo 1989, paraxanthine metabolism in healthy volunteers. Cimetidine pretreatment reduced paraxanthine clearance by 30 percent, and allopurinol did not change it (PMID: 2913277).
• Benowitz 1995, 12-person crossover of oral caffeine or paraxanthine at 2 or 4 mg/kg against placebo. Both raised diastolic blood pressure, plasma epinephrine and free fatty acids, with similar responses at 4 mg/kg (PMID: 8529334).

Animal studies

• Rats, National Institute on Drug Abuse. Paraxanthine produced stronger locomotor activation than caffeine and raised striatal cGMP and dopamine (PMID: 23261866).
• Mice, Stanford University. Paraxanthine promoted wakefulness more strongly and for longer than equimolar caffeine in normal and narcoleptic mice (PMID: 20614853).
• Mice and rats, supplier-authored. In mice, 4 weeks of oral paraxanthine increased grip strength by 17 percent and treadmill performance by 39 percent compared with control (PMID: 35215543). In rats, paraxanthine improved water maze escape times and raised brain BDNF more than caffeine (PMID: 39617850). A 2025 correction to that paper fixed an author's name, three swapped figure captions and the Figure 1 timeline diagram, and listed no change to the results (PMID: 39907818).

Toxicology

• Rats, 2021. No mutagenicity or genotoxicity, acute oral LD50 1,601 mg/kg, and no adverse effects in 28-day and 90-day studies at up to 300 mg/kg a day (PMID: 34597720).
• Rats, 2023. No mutagenicity or genotoxicity, acute oral LD50 829.20 mg/kg, and a 90-day no-observed-adverse-effect level of 185 mg/kg a day for paraxanthine against 150 mg/kg a day for caffeine (PMID: 36818692).
• Mice, 1986 teratology. Injected paraxanthine at 175 or 300 mg/kg a day in mid-pregnancy caused dose-related malformations (PMID: 3798364).

Regulatory filings

Searches on 20 October 2026 found no product containing paraxanthine in Drugs@FDA and no GRAS notice for paraxanthine in the FDA's GRAS Notice Inventory. A regulations.gov search on the same date found no new dietary ingredient notification for paraxanthine. The NIH Office of Dietary Supplements' Dietary Supplement Label Database lists on-market US supplement labels that contain a paraxanthine ingredient.

What the human studies recorded

In the 7-day dose comparison of 50, 100 and 200 mg, there were no significant differences from placebo in blood chemistry panels or in the frequency or severity of reported side effects (PMID: 34960030). In 12 trained runners, there was no evidence that paraxanthine worsened ratings of stimulant side effects or clinical blood markers (PMID: 38725238). In 21 adults, blood pressure did not change at 100 to 300 mg and heart rate was lower than with placebo at two time points after 100 mg (PMID: 38745415).

Paraxanthine has caffeine-like effects on the body: in 12 people, paraxanthine at 4 mg/kg raised diastolic blood pressure, plasma epinephrine and free fatty acids about as much as the same dose of caffeine (PMID: 8529334).

Sleep

In 14 rowers, conditions containing caffeine were associated with poorer subjective sleep quality, while paraxanthine alone showed more favorable sleep-related outcomes; sleep was measured only by questionnaire (PMID: 41918248).

Animal toxicology

A battery of rat studies of one supplier's synthetic paraxanthine found no mutagenicity or genotoxicity, an acute oral LD50 of 1,601 mg/kg, and no deaths or toxic effects in 28-day and 90-day studies at 75, 150 or 300 mg/kg a day, the highest dose tested; two of the three authors were affiliated with that supplier (PMID: 34597720). A second battery on another supplier's paraxanthine found no genotoxicity or mutagenicity, an acute oral LD50 of 829.20 mg/kg, and no deaths or treatment-related adverse effects at up to 185 mg/kg a day for 90 days, while two rats given caffeine at 185 mg/kg a day died (PMID: 36818692).

In pregnant mice given paraxanthine by injection at 175 or 300 mg/kg a day on days 11 and 12 of gestation, malformations rose with dose, mainly cleft palate and limb defects, with a 46 percent malformation rate at 300 mg/kg; the authors judged paraxanthine slightly less toxic to the embryo than caffeine (PMID: 3798364).

Gaps in the safety record

The human paraxanthine trials summarized here lasted from a single dose to 7 days of daily use, so they do not describe long-term use. No human data on paraxanthine supplements in pregnancy were found. In a study that used serum paraxanthine as a marker of caffeine intake in pregnancy, only the highest levels (above 1,845 ng/mL) were linked to spontaneous abortion (adjusted odds ratio 1.9) (PMID: 10572151).